This presentation critically examines the use of network analysis in digital humanities, focusing specifically on Ancient Greek comedy. While platforms like DraCor have advanced this field, our research highlights significant methodological issues, such as segmentation based on large structural units and the misrepresentation of unique dramatic events like the splitting and merging of the chorus. To address these issues, we introduce an experimental annotation schema and a new analytical tool. This tool, developed with the assistance of large language models, serves as a proof-of-concept. It enables a more nuanced, performance-oriented segmentation of the text, moving beyond simple co-presence to model discrete “stage events.” The tool’s visualizations are presented not as objective truths but as a means to discuss fundamental questions about the epistemic value of data. Ultimately, this analysis reveals the tension between computational precision and humanistic interpretation. We highlight the fact that the value of digital tools lies not in their perceived objectivity but in their ability to make interpretative assumptions explicit. The goal is to create tools that facilitate critical scholarly judgment, enhancing rather than replacing humanistic understanding of dramatic texts.